Color Managing Printers

In the process of color management, the single most tricky aspect is dealing with the printer. This article describes the issues involved, and illustrates the use of printer color management using the Spyder3Print from Datacolor.

Spyder3 Elite is a device for calibrating and color managing monitors. The two devices for both monitor and printer are packaged together in the Spyder3 Studio, and come with a sturdy metal carrying case – they are both sensitive measuring devices, after all. This package has a recommended price of Rs. 52,000.It will be assumed in what follows that your display has already been calibrated properly. You can also calibrate your camera, but this is not normally done as lighting conditions and color balance vary so much. Many digital cameras default to a color space defined as sRGB, and more advanced cameras allow also for the use of Adobe RGB. The information about which is used is embedded within the image file that is created.

So, what does all this mean? If you take a photo of a patch of color that is known to be primary red, this should be represented in your computer image as the RGB (red green blue) value of 255,0,0. When that color value is then sent to a screen for display or to a printer, the color that results should also be primary red.

There are two main ways in which the digital values and the actual colors imaged or reproduced do not match up: the range of colors – known as the gamut – that can be represented by any device varies, and the accuracy of matching colors to RGB values is off. The way in which this is managed is by the use of profiles. Basically, a profile is simply a file (with extension .ICC or .ICM) that contains a numerical description of the range of colors that a device can handle. It gives the range of colors, together with the numbers – RGB values – that are associated with them. The profile describes the device’s color gamut, and the associated numerical values.

Using the profile, software can then make adjustments accordingly. To give an over-simplified example, if its profile makes it clear that a particular scanner gives RGB values in which the red is over-emphasized, the software reading the file, something such as Photoshop, perhaps, can adjust the values it receives – it corrects the values that it stores for the peculiarities of the scanner.

When that software then displays the image on screen, assuming that the display device has also been calibrated, it will use the screen profile to adjust the values that it sends to the display. Keeping to a simple example, if the display profile indicates that the display tends to create blues that are too weak, the software will send to the display blue values that are increased by an appropriate amount. A similar process will apply when sending the image to a printer.

By making these adjustments based on the devices’ profiles, the image that is displayed or printed comes much closer to representing the same colors as in the original scanned subject – instead of having reds that are too strong due to the idiosyncrasies of the scanner, and blues too weak thanks to the display.

Of course, it is more complicated than this. The relationship of the numerical values to the range of colors is not simply one such as “over-doing the reds”. The relationship is described either in terms of steps or algebraically. But the biggest problem lies when the gamut of one device is considerably different from another. we’ll come back to this point.

The approach taken a few years ago with Windows and general computing devices was really to reduce everything to a lowest common denominator. This meant assuming a rather narrow color space that just about every device could be expected to cope with, and the use of generic profiles for devices that expressed colors in terms of that color space. This color space is sRGB, properly known as sRGB IEC61966-2.1, and was based on the typical characteristics of a cathode ray monitor of a few years ago. This was devised by Microsoft and Hewlett-Packard and is the expected color space for use on the internet – if you plan to publish any images on the web, always convert them to sRGB before doing so; this should ensure the best results on the widest range of viewing equipment. Typical digital cameras use sRGB by default – but it is worth checking.

It is generally considered that the sRGB color space includes about a third of all possible real-world visible colors – although the definition of the maximum possible number of colors varies. A third may not seem very much, but it is certainly good enough for most purposes – except professional graphics use, and that is where Adobe RGB came in. This is a larger color space than sRGB, and was created by Adobe – the developers of graphics software such as Photoshop – to match the color space that could be reproduced my most color printers – a little less than double the number of colors of sRGB. Adobe sRGB is available for use in many higher-end digital cameras, and should be used if you plan on printing your images.

Using these two color spaces, sRGB for use on general displays on the internet and Adobe RGB for color printing, one can gain reasonably satisfactory results without the use of any calibration or specific profiles. But even these two can show up considerable differences, and one should check which is in use at any time. Consistency is vital. For example, load an image into Photoshop and then switch between color spaces – the difference will be immediately obvious, and one image will be more true-to-life than the other.

However, hardware is improving with time, with camera sensors becoming more sensitive, monitors gaining much greater contrast range, and the gamut of printers increasing. These improvements, which show no signs of slowing down, give good reason for looking into more sophisticated forms of color management than simply relying on sRGB and Adobe RGB.

As mentioned before, one major problem lies when you move from a device with one gamut, to a device with a smaller gamut. This is typically the case when printing. In a PC, an image file is held in RGB format. On your display, the varying values of red, green and blue combine to give the full range of colors. But printers work subtractively – you print onto a white surface, and subtract color from it. The color information is converted into a CMYK color space – cyan, magenta and yellow – together with black. This subtractive process results in the printer gamut being smaller than other typical devices.

This is why Adobe took this color space for the definition of Adobe RGB – its lowest common denominator. If you set your camera to use the Adobe color space, a good quality PC and monitor can easily handle this, and little conversion is needed when printing – except from RGB to CMYK. This is a good choice for amateur photographers to print out their photos.

But the next step up entails some complications, and choices that need to be made. Perhaps the most important entails the concept of mapping “intent”. There are three main methods, called colorimetric, perceptual, and saturation. (We shall not cover saturation, which is most useful when going from a small color space to a larger one.) Of the first of these, there are two variants, known as absolute and relative. The easiest to describe first is perceptual intent.

With perceptual intent, the color range from the larger source gamut is scaled down to match the smaller target color space. The “distance” between the colors is reduced, but their relative positions remain essentially the same. This method gives the best results from the point of view of human perception – hence the name. The two types of colorimetric intent handle the colors that are shared in both gamuts by mapping them directly – actually, there is a difference in the two variants here regarding this, with relative colorimetric not doing a direct mapping, but this need not concern us for now.

The important feature lies with those colors in the source gamut that lie outside of the target gamut – they are all simply mapped to the nearest color point within the target gamut; what one might think of as the edge of that gamut. In this way several colors in the source color space are all mapped to the same color in the target. It is important to consider the colors that are actually in use in a particular image – many cameras and imaging software give different ways of displaying this information, and it is worth taking the time to understand this information.

If, for example, the great majority of the colors in an image lie within the gamut of the target printer device, then using absolute colorimetric intent will be the most accurate – there is no point in scaling them down if they already fit, as you would do with perceptual intent. However, the greater the difference between the source and target gamuts, then the more important become those colors that lie outside of the target gamut. Think of an outdoor shot with much detail in green foliage, many different shades of dark green – colors that might well be outside of the printer gamut. Using colorimetric intent with this image would likely have most of those colors end up the same, losing all the detail. In that instance, perceptual intent would preserve more detail and yield a more satisfying result.

So, having described the main issues involved with color management, and particularly with printers, we will now look at a practical example using the Spyder3Print.

Using the Spyder3Print

The main component of the Spyder3Print is the colorimeter that is used for taking measurements of color patches that have been printed. There is also a guide that you lay down on the paper under inspection, a stand into which the device sits and also a CD containing software. The colorimeter is shaped rather like a large mouse, and the pointed end contains the sensor – just a few mms across – that is used for taking color measurements. Unfortunately, there is zero documentation, and the user is expected touse the help system to learn how to use the device. This does not contain enough much information – not by a long way.

The software itself does not give the user much information either – we also found with the Spyder for calibrating monitors that the data derived by the system is not really accessible to the user. You can see what the developers are trying to do – make the system easy to use and not confuse things with too much information; but this is a device for enthusiastic amateurs and professionals – just the kind of people who are likely to want at least the ability to dig deeper under the hood.

The process of calibrating a printer contains many details, but it is simple to describe overall. You print out a test page that contains a large number of defined color patches, and then each of these is read by the Spyder colorimeter. When all have been read, the software compares the expected colors with those that have been detected, and then proceeds to create a profile for use when printing. On the Windows XP system we used, the profile was placed in the windows\system32\spool\drivers\color\ directory. (For some bizarre reason, different version of Windows use other directories.) From there it is readily found by programs such as Photoshop. And now for the details.

One important issue with calibrating a printer is that if you can, turn off any color management that the printer driver itself might otherwise be trying to do. It is not always easy to check this, and it is not always possible to turn it off. For trying out the Spyder, we used an HP Photosmart C8188 All-in-One. Its printer driver gives two options under color management – Adobe RGB or ColorSmart/sRGB. We used Adobe RGB as this has the largest gamut. In a recent document, HP has described that a new, third, option is being added to its printer drivers: “Managed by application”. This turns off printer driver-based color management, and the printer will not make any color conversions to the image data. This of course assumes that you will be using software that will be able to handle the color management, using the profile that you create with the Spyder or other colorimeter. If you are going to buy a color printer, make sure that the color management can be turned off.

Needless to say, not just the driver color management, but also any other printer driver settings need to be the same when you perform the calibration and create the profile as when you later use the printer. If you intend to use different paper and ink combinations, and want accurate color with these, each combination will need to be profiled separately.

The next step is to run the Spyder software and calibrate the colorimeter. The small stand in which the Spyder can sit when not being used contains a small white patch under the Spyder sensor. When you start the software, you check that it can see the colorimeter, seated in its stand, and then select a menu option for calibration. Of course, it makes more sense to calibrate using the paper that will be used in printing. I have been told this will be particularly beneficial when there are high levels of optical brighteners in the paper, but it seems better to do this all the time.

After noting details regarding the printer and media to be used, you are lead through the process of checking the printer’s quality and media settings; you then come to print out the print sample that will be used for calibration. There are five from which you can choose:

1. Fast target – 150 patches
2. High quality target – 225 patches
3. Expert target – 729 patches (on three sheets, small paper)
4. Expert target – 729 patches (on three sheets, large paper)
5. Extended Greys – for black and white printing

For our testing, we used the fast target with 150 color patches. Time did not allow us to compare the results from using the 150 patch calibration against the 225 or 729 patch methods. There are other factors coming later that perhaps might make a great difference.

When you perform the actual calibration, you place the printed test sheet on a table, and on top of this the guide – this helps you position the Spyder sensor over each patch. You start in the top left corner, and progress along each row. A cursor in the shape of a red triangle on screen shows which patch is being measured, and once the measurement has been take – either by pressing the button on the Spyder or by hitting Enter – the software makes a sound to indicate the measurement, and then another as it automatically moves the cursor along to the next patch.

The user has to take care to move the sensor itself, and make sure it is correctly positioned over each patch. It is easy to keep to the correct patch in the sequence, but care must be taken to ensure the positioning is correct and that the sensor is flat on the paper – otherwise bad measurements will result. It took just over ten minutes to measure all 225 patches in the Fast target – it is well worth doing this slowly and very carefully, although you can go back and retake measurements, if needed. We tried this a few times, and found that measurements were always varying slightly.

The next screen (called “Build Profile Setup”) gives a set of further options for adjustment before the actual profile is built. We’ll come back to this. You enter a name for the file, select any adjustments as necessary, and then press “Next”. The profile is built and saved to disk. You then come to a screen named “SpyderProof”. This allows you to print and compare the results. You can here choose the rendering intent, print the sample, and compare with the image on screen, assuming you have a calibrated screen. It would be a much better idea for actual sample printouts to be provided by Datacolor here – these would enable a much more accurate comparison.

In our tests, we used a standard PDI test file, which certainly seemed to give better results when printed using the created profile than not. However, we also printed some images of artwork, and the results seemed best – most true to the original – when simply using the printer’s built-in Adobe RGB profile. Clearly, this is far from an exact science.

This type of variability is obviously the reason for the existence of the “Build Profile Setup” screen that immediately precedes the creation of a profile. It obviously makes sense the first time you profile your printer to make no other manual adjustments – let the software do its job first, as objectively as possible, and then only later make further manual adjustments.

I asked the distributors of the Spyder why these extra controls are there, and they came back with a mixture of answers, some of which involved personal preferences. The other main points concerned the physical characteristics of the printer and media themselves; these were, in brief:

The presence of optical brighteners in the paper can at times fool the machine and lead to yellowish greys. Matt fine art papers require compensation for shadow detail. Some printers (especially some lasers) are weak in printing highlight details. Profiling tends to work best in printers which have extra grey inks, in other printers you might get some color cast which can be removed manually.

Why bother about color management ?

In an ideal world, if you take a photograph of any scene (or a scan of any image), the colors you see when directly looking at the scene would be indistinguishable from those that you see when that image is displayed on your monitor or printed on your printer. This is clearly critical for high-end applications such as book and magazine publication, but can also be important for semi-professional photography or enthusiastic amateurs.

But it is a far from perfect world, and there is no such thing as a perfect camera, screen, printer, etc.

Not only do the colors not match exactly, but also the range of colors – the gamut – that any device can detect or reproduce will vary from one device to another; and all devices have gamuts that are considerably less than that perceived by human vision.

The process of making adjustments as colors are mapped from one device to another so as to bring the colors displayed or printed as close as possible to those perceived in the original scene is called color management.

Conclusions

Profiling a display with the Spyder Elite is a simple process. After some initial setup and checks, you place the Spyder against the screen, and the software runs through various color patches, taking measurements as it goes. The profile is then created and used by the system. It is a good idea to repeat the calibration every few weeks, but that is basically it, certainly with a modern, good quality display.

Our experience certainly suggests that profiling your printer with a system such as the Datacolor Spyder is worth doing, but much more is needed in terms of user judgement and time. Straightforward objective profiling simply does not work properly. You need to become familiar with your printer, understand the types of printing that you need to do, and how the printer responds to different profile adjustments and media. It is best to treat a tool like the Spyder as an aid in understanding how to adjust your printer profile. If you think that you can run just one calibration and get excellent results, think again, unless, of course, you have stumbled upon that elusive near-perfect, printer-paper-ink combination.

Getting Spyder

Contact: Eesh Dewan
Phone: +91-11-4153 9116
Email: info@bhushanphoto.com
Web site: www.bhushanphoto.com